Sarom Leang

Electronic structure theory (especially quantum chemistry) has thrived and has become increasingly relevant to a broad spectrum of scientific endeavors as the sophistication of both computer architectures and software engineering has advanced. This article provides a brief history of advances in both hardware and software, from the early days of IBM mainframes to the current emphasis on accelerators and modern programming practices.

A discussion of many of the recently implemented features of GAMESS (General Atomic and Molecular Electronic Structure System) and LibCChem (the C++ CPU/GPU library associated with GAMESS) is presented. These features include fragmentation methods such as the fragment molecular orbital, effective fragment potential and effective fragment molecular orbital methods, hybrid MPI/OpenMP approaches to Hartree–Fock, and resolution of the identity second order perturbation theory. Many new coupled… Show more

A discussion of many of the recently implemented features of GAMESS (General Atomic and Molecular Electronic Structure System) and LibCChem (the C++ CPU/GPU library associated with GAMESS) is presented. These features include fragmentation methods such as the fragment molecular orbital, effective fragment potential and effective fragment molecular orbital methods, hybrid MPI/OpenMP approaches to Hartree–Fock, and resolution of the identity second order perturbation theory. Many new coupled cluster theory methods have been implemented in GAMESS, as have multiple levels of density functional/tight binding theory. The role of accelerators, especially graphical processing units, is discussed in the context of the new features of LibCChem, as it is the associated problem of power consumption as the power of computers increases dramatically. The process by which a complex program suite such as GAMESS is maintained and developed is considered. Future developments are briefly summarized. Show less

The computational efficiency and energy-to-solution of several applications using the GAMESS quantum chemistry suite of codes is evaluated for 32-bit and 64-bit ARM-based computers, and compared to an x86 machine. The x86 system completes all benchmark computations more quickly than either ARM system and is the best choice to minimize time to solution. The ARM64 and ARM32 computational performances are similar to each other for Hartree–Fock and density functional theory energy calculations… Show more

The computational efficiency and energy-to-solution of several applications using the GAMESS quantum chemistry suite of codes is evaluated for 32-bit and 64-bit ARM-based computers, and compared to an x86 machine. The x86 system completes all benchmark computations more quickly than either ARM system and is the best choice to minimize time to solution. The ARM64 and ARM32 computational performances are similar to each other for Hartree–Fock and density functional theory energy calculations. However, for memory-intensive second-order perturbation theory energy and gradient computations the lower ARM32 read/write memory bandwidth results in computation times as much as 86% longer than on the ARM64 system. The ARM32 system is more energy efficient than the x86 and ARM64 CPUs for all benchmarked methods, while the ARM64 CPU is more energy efficient than the x86 CPU for some core counts and molecular sizes. Show less

Increasingly, modern computer systems comprise a multicore general-purpose processor augmented with a number of special purpose devices or accelerators connected via an external interface such as a PCI bus. The NVIDIA Kepler Graphical Processing Unit (GPU) and the Intel Phi are two examples of such accelerators. Accelerators offer peak performances that can be well above those of the host processor. How to exploit this heterogeneous environment for legacy application codes is not, however… Show more

Increasingly, modern computer systems comprise a multicore general-purpose processor augmented with a number of special purpose devices or accelerators connected via an external interface such as a PCI bus. The NVIDIA Kepler Graphical Processing Unit (GPU) and the Intel Phi are two examples of such accelerators. Accelerators offer peak performances that can be well above those of the host processor. How to exploit this heterogeneous environment for legacy application codes is not, however, straightforward. This paper considers how matrix operations in typical quantum chemical calculations can be migrated to the GPU and Phi systems. Double precision general matrix multiply operations are endemic in electronic structure calculations, especially methods that include electron correlation, such as density functional theory, second order perturbation theory, and coupled cluster theory. The use of approaches that automatically determine whether to use the host or an accelerator, based on problem size, is explored, with computations that are occurring on the accelerator and/or the host. For data-transfers over PCI-e, the GPU provides the best overall performance for data sizes up to 4096 MB with consistent upload and download rates between 5–5.6 GB/s and 5.4–6.3 GB/s, respectively. The GPU outperforms the Phi for both square and nonsquare matrix multiplications. Show less

Meta-generalized gradient approximation (meta-GGA) exchange-correlation density functionals depend on the Kohn-Sham (KS) orbitals through the kinetic energy density. The KS orbitals in turn depend functionally on the electron density. However, the functional dependence of the KS orbitals is indirect, i.e., not given by an explicit expression, and the computation of analytic functional derivatives of meta-GGA functionals with respect to the density imposes a challenge. The practical solution… Show more

Meta-generalized gradient approximation (meta-GGA) exchange-correlation density functionals depend on the Kohn-Sham (KS) orbitals through the kinetic energy density. The KS orbitals in turn depend functionally on the electron density. However, the functional dependence of the KS orbitals is indirect, i.e., not given by an explicit expression, and the computation of analytic functional derivatives of meta-GGA functionals with respect to the density imposes a challenge. The practical solution used in many computer implementations of meta-GGA density functionals for ground-state calculations is abstracted and generalized to a class of density functionals that is broader than meta-GGAs and to any order of functional differentiation. Importantly, the TDDFT working equations for meta-GGA density functionals are presented here for the first time, together with the technical details of their computer implementation. The analysis presented here also uncovers the implicit assumptions in the practical solution to computing functional derivatives of meta-GGA density functionals. The connection between the approximation that is invoked in taking functional derivatives of density functionals, the non-uniqueness with respect to the KS orbitals, and the non-locality of the resultant potential is also discussed. Show less

Water and argon hexamers are examined using correlated wavefunction methods, the effective fragment potential (EFP) method, and Hartree–Fock (HF) theory and several density functional theory (DFT) functionals. The HF and DFT methods have been employed both with and without dispersion corrections. The computationally inexpensive EFP method captures the high-level coupled cluster binding energy and relative isomer energy predictions very well for both types of hexamer, much better than the DFT… Show more

Water and argon hexamers are examined using correlated wavefunction methods, the effective fragment potential (EFP) method, and Hartree–Fock (HF) theory and several density functional theory (DFT) functionals. The HF and DFT methods have been employed both with and without dispersion corrections. The computationally inexpensive EFP method captures the high-level coupled cluster binding energy and relative isomer energy predictions very well for both types of hexamer, much better than the DFT methods (DFT-D), even those that include dispersion corrections. Interestingly, the dispersion corrected HF method (HF-D) does very well. When the DFT-D methods perform reasonably, they do so because of a fortuitous off-setting cancellation of errors two-body and many-body contributions to the binding energy. The HF-D method does not rely on such error cancellation, while the EFP method captures both two-body and many-body contributions to the water hexamer very well. The many body contribution to the argon cluster are small and are most likely due to dispersion. Show less

Ab initio calculations were carried out to study the hydrolysis mechanism of 1-substituted silatranes in the presence of an acid (acid-catalyzed) and an additional water (water-assisted). Compared with the neutral hydrolysis mechanism involving one water, use of an acid catalyst reduces the barrier associated with the rate-limiting step by ≈14 kcal/mol. A modest decrease of ≈5 kcal/mol is predicted when an additional water molecule is added to the neutral hydrolysis mechanism involving one… Show more

Ab initio calculations were carried out to study the hydrolysis mechanism of 1-substituted silatranes in the presence of an acid (acid-catalyzed) and an additional water (water-assisted). Compared with the neutral hydrolysis mechanism involving one water, use of an acid catalyst reduces the barrier associated with the rate-limiting step by ≈14 kcal/mol. A modest decrease of ≈5 kcal/mol is predicted when an additional water molecule is added to the neutral hydrolysis mechanism involving one water. The combination of an acid catalyst and an additional water molecule reduces the barrier by ≈ 27 kcal/mol. Bond order analysis suggests ring cleavage involving the bond breaking of a siloxane and silanol group during the neutral and acid-catalyzed hydrolysis of 1-substituted silatranes, respectively. Solvent effects, represented by the PCM continuum model, do not qualitatively alter computational gas-phase results. Show less

The performance of 24 density functionals, including 14 meta-generalized gradient approximation (mGGA) functionals, is assessed for the calculation of vertical excitation energies against an experimental benchmark set comprising 14 small- to medium-sized compounds with 101 total excited states. The experimental benchmark set consists of singlet, triplet, valence, and Rydberg excited states. The global-hybrid (GH) version of the Perdew-Burke-Ernzerhoff GGA density functional (PBE0) is found to… Show more

The performance of 24 density functionals, including 14 meta-generalized gradient approximation (mGGA) functionals, is assessed for the calculation of vertical excitation energies against an experimental benchmark set comprising 14 small- to medium-sized compounds with 101 total excited states. The experimental benchmark set consists of singlet, triplet, valence, and Rydberg excited states. The global-hybrid (GH) version of the Perdew-Burke-Ernzerhoff GGA density functional (PBE0) is found to offer the best overall performance with a mean absolute error (MAE) of 0.28 eV. The GH-mGGA Minnesota 2006 density functional with 54% Hartree-Fock exchange (M06-2X) gives a lower MAE of 0.26 eV, but this functional encounters some convergence problems in the ground state. The local density approximation functional consisting of the Slater exchange and Volk-Wilk-Nusair correlation functional (SVWN) outperformed all non-GH GGAs tested. The best pure density functional performance is obtained with the local version of the Minnesota 2006 mGGA density functional (M06-L) with an MAE of 0.41 eV.

Relative to the Journal of Chemical Physics:
► Most cited article published in 2012 #1
► Most downloaded March 2012 #1
► Most accessed article Q1 2013 #3
► Most downloaded March 2013 #8
► Most downloaded May 2013 #7 Show less

The combined time-dependent density functional theory effective fragment potential method (TDDFT/EFP1) is applied to a study of the solvent-induced shift of the lowest singlet π → π* charge-transfer excited state of p-nitroaniline (pNA) from the gas to the condensed phase in water. Molecular dynamics simulations of pNA with 150 EFP1 water molecules are used to model the condensed-phase and generate a simulated spectrum of the lowest singlet charge-transfer excitation. The TDDFT/EFP1 method… Show more

The combined time-dependent density functional theory effective fragment potential method (TDDFT/EFP1) is applied to a study of the solvent-induced shift of the lowest singlet π → π* charge-transfer excited state of p-nitroaniline (pNA) from the gas to the condensed phase in water. Molecular dynamics simulations of pNA with 150 EFP1 water molecules are used to model the condensed-phase and generate a simulated spectrum of the lowest singlet charge-transfer excitation. The TDDFT/EFP1 method successfully reproduces the experimental condensed-phase π → π* vertical excitation energy and solvent-induced red shift of pNA in water. The largest contribution to the red shift comes from Coulomb interactions, between pNA and water, and solute relaxation. The solvent shift contributions reflect the increase in zwitterionic character of pNA upon solvation. Show less

Decreasing crude oil supplies has lead to the import and use of foreign fuels. Middle distillate fuels are used for ground, water, and air transportation purposes. It is thus important to study the chemical composition of foreign fuels to elucidate the quality of middle distillate fuels in terms of fuel instability. Instability and incompatibility are terms that denote fuel degradation. Reaction mechanisms leading to fuel degradation are difficult to identify. Results from our laboratory… Show more

Decreasing crude oil supplies has lead to the import and use of foreign fuels. Middle distillate fuels are used for ground, water, and air transportation purposes. It is thus important to study the chemical composition of foreign fuels to elucidate the quality of middle distillate fuels in terms of fuel instability. Instability and incompatibility are terms that denote fuel degradation. Reaction mechanisms leading to fuel degradation are difficult to identify. Results from our laboratory indicate that organo-nitrogen compounds are active participants in this sedimentation process. Middle distillate fuel from Thailand purported to form sediment and thus be unstable was analyzed.

Print copies are available upon request. Show less